Add files using upload-large-folder tool

lib/python3.12/site-packages/torch/include/ATen/ops/_adaptive_avg_pool3d_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,26 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API at::Tensor & _adaptive_avg_pool3d_out(at::Tensor & out, const at::Tensor & self, at::IntArrayRef output_size);
+TORCH_API at::Tensor & _adaptive_avg_pool3d_outf(const at::Tensor & self, at::IntArrayRef output_size, at::Tensor & out);
+TORCH_API at::Tensor & _adaptive_avg_pool3d_symint_out(at::Tensor & out, const at::Tensor & self, c10::SymIntArrayRef output_size);
+TORCH_API at::Tensor & _adaptive_avg_pool3d_symint_outf(const at::Tensor & self, c10::SymIntArrayRef output_size, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_add_batch_dim_ops.h

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _add_batch_dim {
+  using schema = at::Tensor (const at::Tensor &, int64_t, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_add_batch_dim";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_add_batch_dim(Tensor self, int batch_dim, int level) -> Tensor";
+  static at::Tensor call(const at::Tensor & self, int64_t batch_dim, int64_t level);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, int64_t batch_dim, int64_t level);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_amp_update_scale_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _amp_update_scale_ {
+  using schema = at::Tensor & (at::Tensor &, at::Tensor &, const at::Tensor &, double, double, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_amp_update_scale_";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_amp_update_scale_(Tensor(a!) self, Tensor(b!) growth_tracker, Tensor found_inf, float scale_growth_factor, float scale_backoff_factor, int growth_interval) -> Tensor(a!)";
+  static at::Tensor & call(at::Tensor & self, at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, at::Tensor & self, at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval);
+};
+
+struct TORCH_API _amp_update_scale_out {
+  using schema = at::Tensor & (const at::Tensor &, at::Tensor &, const at::Tensor &, double, double, int64_t, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_amp_update_scale";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_amp_update_scale.out(Tensor self, Tensor(b!) growth_tracker, Tensor found_inf, float scale_growth_factor, float scale_backoff_factor, int growth_interval, *, Tensor(a!) out) -> Tensor(a!)";
+  static at::Tensor & call(const at::Tensor & self, at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval, at::Tensor & out);
+};
+
+struct TORCH_API _amp_update_scale {
+  using schema = ::std::tuple<at::Tensor,at::Tensor> (const at::Tensor &, const at::Tensor &, const at::Tensor &, double, double, int64_t);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_amp_update_scale";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_amp_update_scale(Tensor self, Tensor growth_tracker, Tensor found_inf, float scale_growth_factor, float scale_backoff_factor, int growth_interval) -> (Tensor, Tensor growth_tracker_out)";
+  static ::std::tuple<at::Tensor,at::Tensor> call(const at::Tensor & self, const at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval);
+  static ::std::tuple<at::Tensor,at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, const at::Tensor & growth_tracker, const at::Tensor & found_inf, double scale_growth_factor, double scale_backoff_factor, int64_t growth_interval);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_assert_async_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API void _assert_async(const at::Tensor & self);
+TORCH_API void _assert_async(const at::Tensor & self, c10::string_view assert_msg);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_autocast_to_reduced_precision_compositeimplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeimplicitautograd {
+
+TORCH_API at::Tensor _autocast_to_reduced_precision(const at::Tensor & self, bool cuda_enabled, bool cpu_enabled, at::ScalarType cuda_dtype, at::ScalarType cpu_dtype);
+
+} // namespace compositeimplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_batch_norm_impl_index_backward_compositeimplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeimplicitautograd {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _batch_norm_impl_index_backward(int64_t impl_index, const at::Tensor & input, const at::Tensor & grad_output, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, const ::std::optional<at::Tensor> & save_mean, const ::std::optional<at::Tensor> & save_var_transform, bool train, double eps, ::std::array<bool,3> output_mask, const at::Tensor & reservedSpace);
+
+} // namespace compositeimplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_batch_norm_impl_index_native.h

@@ -0,0 +1,21 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor,at::Tensor,int64_t> _batch_norm_impl_index(const at::Tensor & input, const ::std::optional<at::Tensor> & weight, const ::std::optional<at::Tensor> & bias, const ::std::optional<at::Tensor> & running_mean, const ::std::optional<at::Tensor> & running_var, bool training, double momentum, double eps, bool cudnn_enabled);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cast_Double_native.h

@@ -0,0 +1,21 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor _cast_Double(const at::Tensor & self, bool non_blocking=false);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cast_Float_compositeimplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeimplicitautograd {
+
+TORCH_API at::Tensor _cast_Float(const at::Tensor & self, bool non_blocking=false);
+
+} // namespace compositeimplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cast_Long.h

@@ -0,0 +1,30 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_cast_Long_ops.h>
+
+namespace at {
+
+
+// aten::_cast_Long(Tensor self, bool non_blocking=False) -> Tensor
+inline at::Tensor _cast_Long(const at::Tensor & self, bool non_blocking=false) {
+    return at::_ops::_cast_Long::call(self, non_blocking);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_cast_Short_native.h

@@ -0,0 +1,21 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor _cast_Short(const at::Tensor & self, bool non_blocking=false);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_conj_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API at::Tensor _conj(const at::Tensor & self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_convolution_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _convolution {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, c10::SymIntArrayRef, c10::SymIntArrayRef, c10::SymIntArrayRef, bool, c10::SymIntArrayRef, c10::SymInt, bool, bool, bool, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_convolution";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_convolution(Tensor input, Tensor weight, Tensor? bias, SymInt[] stride, SymInt[] padding, SymInt[] dilation, bool transposed, SymInt[] output_padding, SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled, bool allow_tf32) -> Tensor";
+  static at::Tensor call(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, bool transposed, c10::SymIntArrayRef output_padding, c10::SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled, bool allow_tf32);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, bool transposed, c10::SymIntArrayRef output_padding, c10::SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled, bool allow_tf32);
+};
+
+struct TORCH_API _convolution_deprecated {
+  using schema = at::Tensor (const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, c10::SymIntArrayRef, c10::SymIntArrayRef, c10::SymIntArrayRef, bool, at::IntArrayRef, c10::SymInt, bool, bool, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_convolution";
+  static constexpr const char* overload_name = "deprecated";
+  static constexpr const char* schema_str = "_convolution.deprecated(Tensor input, Tensor weight, Tensor? bias, SymInt[] stride, SymInt[] padding, SymInt[] dilation, bool transposed, int[] output_padding, SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled) -> Tensor";
+  static at::Tensor call(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, bool transposed, at::IntArrayRef output_padding, c10::SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, bool transposed, at::IntArrayRef output_padding, c10::SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled);
+};
+
+struct TORCH_API _convolution_out {
+  using schema = at::Tensor & (const at::Tensor &, const at::Tensor &, const ::std::optional<at::Tensor> &, c10::SymIntArrayRef, c10::SymIntArrayRef, c10::SymIntArrayRef, bool, c10::SymIntArrayRef, c10::SymInt, bool, bool, bool, bool, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_convolution";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_convolution.out(Tensor input, Tensor weight, Tensor? bias, SymInt[] stride, SymInt[] padding, SymInt[] dilation, bool transposed, SymInt[] output_padding, SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled, bool allow_tf32, *, Tensor(a!) out) -> Tensor(a!)";
+  static at::Tensor & call(const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, bool transposed, c10::SymIntArrayRef output_padding, c10::SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled, bool allow_tf32, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & input, const at::Tensor & weight, const ::std::optional<at::Tensor> & bias, c10::SymIntArrayRef stride, c10::SymIntArrayRef padding, c10::SymIntArrayRef dilation, bool transposed, c10::SymIntArrayRef output_padding, c10::SymInt groups, bool benchmark, bool deterministic, bool cudnn_enabled, bool allow_tf32, at::Tensor & out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_copy_from_and_resize_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API at::Tensor & _copy_from_and_resize_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & dst);
+TORCH_API at::Tensor & _copy_from_and_resize_outf(const at::Tensor & self, const at::Tensor & dst, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cslt_sparse_mm_search_native.h

@@ -0,0 +1,21 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API int64_t _cslt_sparse_mm_search(const at::Tensor & compressed_A, const at::Tensor & dense_B, const ::std::optional<at::Tensor> & bias={}, const ::std::optional<at::Tensor> & alpha={}, ::std::optional<at::ScalarType> out_dtype=::std::nullopt, bool transpose_result=false);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_ctc_loss.h

@@ -0,0 +1,53 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_ctc_loss_ops.h>
+
+namespace at {
+
+
+// aten::_ctc_loss(Tensor log_probs, Tensor targets, int[] input_lengths, int[] target_lengths, int blank=0, bool zero_infinity=False) -> (Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor> _ctc_loss(const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, int64_t blank=0, bool zero_infinity=false) {
+    return at::_ops::_ctc_loss::call(log_probs, targets, input_lengths, target_lengths, blank, zero_infinity);
+}
+
+// aten::_ctc_loss.Tensor(Tensor log_probs, Tensor targets, Tensor input_lengths, Tensor target_lengths, int blank=0, bool zero_infinity=False) -> (Tensor, Tensor)
+inline ::std::tuple<at::Tensor,at::Tensor> _ctc_loss(const at::Tensor & log_probs, const at::Tensor & targets, const at::Tensor & input_lengths, const at::Tensor & target_lengths, int64_t blank=0, bool zero_infinity=false) {
+    return at::_ops::_ctc_loss_Tensor::call(log_probs, targets, input_lengths, target_lengths, blank, zero_infinity);
+}
+
+// aten::_ctc_loss.out(Tensor log_probs, Tensor targets, int[] input_lengths, int[] target_lengths, int blank=0, bool zero_infinity=False, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> _ctc_loss_out(at::Tensor & out0, at::Tensor & out1, const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, int64_t blank=0, bool zero_infinity=false) {
+    return at::_ops::_ctc_loss_out::call(log_probs, targets, input_lengths, target_lengths, blank, zero_infinity, out0, out1);
+}
+// aten::_ctc_loss.out(Tensor log_probs, Tensor targets, int[] input_lengths, int[] target_lengths, int blank=0, bool zero_infinity=False, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> _ctc_loss_outf(const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, int64_t blank, bool zero_infinity, at::Tensor & out0, at::Tensor & out1) {
+    return at::_ops::_ctc_loss_out::call(log_probs, targets, input_lengths, target_lengths, blank, zero_infinity, out0, out1);
+}
+
+// aten::_ctc_loss.Tensor_out(Tensor log_probs, Tensor targets, Tensor input_lengths, Tensor target_lengths, int blank=0, bool zero_infinity=False, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> _ctc_loss_out(at::Tensor & out0, at::Tensor & out1, const at::Tensor & log_probs, const at::Tensor & targets, const at::Tensor & input_lengths, const at::Tensor & target_lengths, int64_t blank=0, bool zero_infinity=false) {
+    return at::_ops::_ctc_loss_Tensor_out::call(log_probs, targets, input_lengths, target_lengths, blank, zero_infinity, out0, out1);
+}
+// aten::_ctc_loss.Tensor_out(Tensor log_probs, Tensor targets, Tensor input_lengths, Tensor target_lengths, int blank=0, bool zero_infinity=False, *, Tensor(a!) out0, Tensor(b!) out1) -> (Tensor(a!), Tensor(b!))
+inline ::std::tuple<at::Tensor &,at::Tensor &> _ctc_loss_outf(const at::Tensor & log_probs, const at::Tensor & targets, const at::Tensor & input_lengths, const at::Tensor & target_lengths, int64_t blank, bool zero_infinity, at::Tensor & out0, at::Tensor & out1) {
+    return at::_ops::_ctc_loss_Tensor_out::call(log_probs, targets, input_lengths, target_lengths, blank, zero_infinity, out0, out1);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_ctc_loss_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor> _ctc_loss(const at::Tensor & log_probs, const at::Tensor & targets, at::IntArrayRef input_lengths, at::IntArrayRef target_lengths, int64_t blank=0, bool zero_infinity=false);
+TORCH_API ::std::tuple<at::Tensor,at::Tensor> _ctc_loss(const at::Tensor & log_probs, const at::Tensor & targets, const at::Tensor & input_lengths, const at::Tensor & target_lengths, int64_t blank=0, bool zero_infinity=false);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_cudnn_rnn_flatten_weight_native.h

@@ -0,0 +1,22 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API at::Tensor & _cudnn_rnn_flatten_weight_out_symint(at::TensorList weight_arr, int64_t weight_stride0, c10::SymInt input_size, int64_t mode, c10::SymInt hidden_size, c10::SymInt proj_size, int64_t num_layers, bool batch_first, bool bidirectional, at::Tensor & out);
+TORCH_API at::Tensor _cudnn_rnn_flatten_weight(at::TensorList weight_arr, int64_t weight_stride0, int64_t input_size, int64_t mode, int64_t hidden_size, int64_t proj_size, int64_t num_layers, bool batch_first, bool bidirectional);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_embedding_bag_backward.h

@@ -0,0 +1,47 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_embedding_bag_backward_ops.h>
+
+namespace at {
+
+
+// aten::_embedding_bag_backward(Tensor grad, Tensor indices, Tensor offsets, Tensor offset2bag, Tensor bag_size, Tensor maximum_indices, SymInt num_weights, bool scale_grad_by_freq, int mode, bool sparse, Tensor? per_sample_weights, int padding_idx=-1) -> Tensor
+inline at::Tensor _embedding_bag_backward(const at::Tensor & grad, const at::Tensor & indices, const at::Tensor & offsets, const at::Tensor & offset2bag, const at::Tensor & bag_size, const at::Tensor & maximum_indices, int64_t num_weights, bool scale_grad_by_freq, int64_t mode, bool sparse, const ::std::optional<at::Tensor> & per_sample_weights, int64_t padding_idx=-1) {
+    return at::_ops::_embedding_bag_backward::call(grad, indices, offsets, offset2bag, bag_size, maximum_indices, num_weights, scale_grad_by_freq, mode, sparse, per_sample_weights, padding_idx);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same_v<T, int64_t>>>
+  at::Tensor _embedding_bag_backward(const at::Tensor & grad, const at::Tensor & indices, const at::Tensor & offsets, const at::Tensor & offset2bag, const at::Tensor & bag_size, const at::Tensor & maximum_indices, int64_t num_weights, bool scale_grad_by_freq, int64_t mode, bool sparse, const ::std::optional<at::Tensor> & per_sample_weights, int64_t padding_idx=-1) {
+    return at::_ops::_embedding_bag_backward::call(grad, indices, offsets, offset2bag, bag_size, maximum_indices, num_weights, scale_grad_by_freq, mode, sparse, per_sample_weights, padding_idx);
+  }
+}
+
+// aten::_embedding_bag_backward(Tensor grad, Tensor indices, Tensor offsets, Tensor offset2bag, Tensor bag_size, Tensor maximum_indices, SymInt num_weights, bool scale_grad_by_freq, int mode, bool sparse, Tensor? per_sample_weights, int padding_idx=-1) -> Tensor
+inline at::Tensor _embedding_bag_backward_symint(const at::Tensor & grad, const at::Tensor & indices, const at::Tensor & offsets, const at::Tensor & offset2bag, const at::Tensor & bag_size, const at::Tensor & maximum_indices, c10::SymInt num_weights, bool scale_grad_by_freq, int64_t mode, bool sparse, const ::std::optional<at::Tensor> & per_sample_weights, int64_t padding_idx=-1) {
+    return at::_ops::_embedding_bag_backward::call(grad, indices, offsets, offset2bag, bag_size, maximum_indices, num_weights, scale_grad_by_freq, mode, sparse, per_sample_weights, padding_idx);
+}
+namespace symint {
+  template <typename T, typename = std::enable_if_t<std::is_same_v<T, c10::SymInt>>>
+  at::Tensor _embedding_bag_backward(const at::Tensor & grad, const at::Tensor & indices, const at::Tensor & offsets, const at::Tensor & offset2bag, const at::Tensor & bag_size, const at::Tensor & maximum_indices, c10::SymInt num_weights, bool scale_grad_by_freq, int64_t mode, bool sparse, const ::std::optional<at::Tensor> & per_sample_weights, int64_t padding_idx=-1) {
+    return at::_ops::_embedding_bag_backward::call(grad, indices, offsets, offset2bag, bag_size, maximum_indices, num_weights, scale_grad_by_freq, mode, sparse, per_sample_weights, padding_idx);
+  }
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_euclidean_dist_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API at::Tensor _euclidean_dist(const at::Tensor & x1, const at::Tensor & x2);
+TORCH_API at::Tensor & _euclidean_dist_out(at::Tensor & out, const at::Tensor & x1, const at::Tensor & x2);
+TORCH_API at::Tensor & _euclidean_dist_outf(const at::Tensor & x1, const at::Tensor & x2, at::Tensor & out);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_fake_quantize_learnable_per_tensor_affine_backward_native.h

@@ -0,0 +1,21 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::tuple<at::Tensor,at::Tensor,at::Tensor> _fake_quantize_learnable_per_tensor_affine_backward(const at::Tensor & grad, const at::Tensor & self, const at::Tensor & scale, const at::Tensor & zero_point, int64_t quant_min, int64_t quant_max, double grad_factor=1.0);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_fft_r2c_cuda_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor _fft_r2c(const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, bool onesided);
+TORCH_API at::Tensor & _fft_r2c_out(at::Tensor & out, const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, bool onesided);
+TORCH_API at::Tensor & _fft_r2c_outf(const at::Tensor & self, at::IntArrayRef dim, int64_t normalization, bool onesided, at::Tensor & out);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_asin_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_asin {
+  using schema = ::std::vector<at::Tensor> (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_asin";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_asin(Tensor[] self) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_asin_ {
+  using schema = void (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_asin_";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_asin_(Tensor(a!)[] self) -> ()";
+  static void call(at::TensorList self);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_asin_out {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_asin";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_foreach_asin.out(Tensor[] self, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_atan_native.h

@@ -0,0 +1,25 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_atan_slow(at::TensorList self);
+TORCH_API void _foreach_atan_out(at::TensorList self, at::TensorList out);
+TORCH_API void foreach_tensor_atan_slow_(at::TensorList self);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_atan_cuda(at::TensorList self);
+TORCH_API void foreach_tensor_atan_cuda_(at::TensorList self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_copy_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,26 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_copy(at::TensorList self, at::TensorList src, bool non_blocking=false);
+TORCH_API void _foreach_copy_out(at::TensorList out, at::TensorList self, at::TensorList src, bool non_blocking=false);
+TORCH_API void _foreach_copy_outf(at::TensorList self, at::TensorList src, bool non_blocking, at::TensorList out);
+TORCH_API void _foreach_copy_(at::TensorList self, at::TensorList src, bool non_blocking=false);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_div_ops.h

@@ -0,0 +1,149 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_div_Scalar {
+  using schema = ::std::vector<at::Tensor> (at::TensorList, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "Scalar";
+  static constexpr const char* schema_str = "_foreach_div.Scalar(Tensor[] self, Scalar scalar) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self, const at::Scalar & scalar);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Scalar & scalar);
+};
+
+struct TORCH_API _foreach_div__Scalar {
+  using schema = void (at::TensorList, const at::Scalar &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div_";
+  static constexpr const char* overload_name = "Scalar";
+  static constexpr const char* schema_str = "_foreach_div_.Scalar(Tensor(a!)[] self, Scalar scalar) -> ()";
+  static void call(at::TensorList self, const at::Scalar & scalar);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Scalar & scalar);
+};
+
+struct TORCH_API _foreach_div_List {
+  using schema = ::std::vector<at::Tensor> (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "List";
+  static constexpr const char* schema_str = "_foreach_div.List(Tensor[] self, Tensor[] other) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self, at::TensorList other);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList other);
+};
+
+struct TORCH_API _foreach_div__List {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div_";
+  static constexpr const char* overload_name = "List";
+  static constexpr const char* schema_str = "_foreach_div_.List(Tensor(a!)[] self, Tensor[] other) -> ()";
+  static void call(at::TensorList self, at::TensorList other);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList other);
+};
+
+struct TORCH_API _foreach_div_ScalarList {
+  using schema = ::std::vector<at::Tensor> (at::TensorList, at::ArrayRef<at::Scalar>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "ScalarList";
+  static constexpr const char* schema_str = "_foreach_div.ScalarList(Tensor[] self, Scalar[] scalars) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+};
+
+struct TORCH_API _foreach_div__ScalarList {
+  using schema = void (at::TensorList, at::ArrayRef<at::Scalar>);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div_";
+  static constexpr const char* overload_name = "ScalarList";
+  static constexpr const char* schema_str = "_foreach_div_.ScalarList(Tensor(a!)[] self, Scalar[] scalars) -> ()";
+  static void call(at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::ArrayRef<at::Scalar> scalars);
+};
+
+struct TORCH_API _foreach_div_Tensor {
+  using schema = ::std::vector<at::Tensor> (at::TensorList, const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "Tensor";
+  static constexpr const char* schema_str = "_foreach_div.Tensor(Tensor[] self, Tensor other) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self, const at::Tensor & other);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Tensor & other);
+};
+
+struct TORCH_API _foreach_div__Tensor {
+  using schema = void (at::TensorList, const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div_";
+  static constexpr const char* overload_name = "Tensor";
+  static constexpr const char* schema_str = "_foreach_div_.Tensor(Tensor(a!)[] self, Tensor other) -> ()";
+  static void call(at::TensorList self, const at::Tensor & other);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Tensor & other);
+};
+
+struct TORCH_API _foreach_div_Scalar_out {
+  using schema = void (at::TensorList, const at::Scalar &, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "Scalar_out";
+  static constexpr const char* schema_str = "_foreach_div.Scalar_out(Tensor[] self, Scalar scalar, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Scalar & scalar, at::TensorList out);
+};
+
+struct TORCH_API _foreach_div_List_out {
+  using schema = void (at::TensorList, at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "List_out";
+  static constexpr const char* schema_str = "_foreach_div.List_out(Tensor[] self, Tensor[] other, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::TensorList other, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList other, at::TensorList out);
+};
+
+struct TORCH_API _foreach_div_ScalarList_out {
+  using schema = void (at::TensorList, at::ArrayRef<at::Scalar>, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "ScalarList_out";
+  static constexpr const char* schema_str = "_foreach_div.ScalarList_out(Tensor[] self, Scalar[] scalars, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::ArrayRef<at::Scalar> scalars, at::TensorList out);
+};
+
+struct TORCH_API _foreach_div_Tensor_out {
+  using schema = void (at::TensorList, const at::Tensor &, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_div";
+  static constexpr const char* overload_name = "Tensor_out";
+  static constexpr const char* schema_str = "_foreach_div.Tensor_out(Tensor[] self, Tensor other, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, const at::Tensor & other, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, const at::Tensor & other, at::TensorList out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_log.h

@@ -0,0 +1,44 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_foreach_log_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_log(Tensor[] self) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_log(at::TensorList self) {
+    return at::_ops::_foreach_log::call(self);
+}
+
+// aten::_foreach_log_(Tensor(a!)[] self) -> ()
+inline void _foreach_log_(at::TensorList self) {
+    return at::_ops::_foreach_log_::call(self);
+}
+
+// aten::_foreach_log.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_log_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_log_out::call(self, out);
+}
+// aten::_foreach_log.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_log_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_log_out::call(self, out);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_log1p.h

@@ -0,0 +1,44 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_foreach_log1p_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_log1p(Tensor[] self) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_log1p(at::TensorList self) {
+    return at::_ops::_foreach_log1p::call(self);
+}
+
+// aten::_foreach_log1p_(Tensor(a!)[] self) -> ()
+inline void _foreach_log1p_(at::TensorList self) {
+    return at::_ops::_foreach_log1p_::call(self);
+}
+
+// aten::_foreach_log1p.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_log1p_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_log1p_out::call(self, out);
+}
+// aten::_foreach_log1p.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_log1p_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_log1p_out::call(self, out);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_log1p_cuda_dispatch.h

@@ -0,0 +1,24 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_log1p(at::TensorList self);
+TORCH_API void _foreach_log1p_(at::TensorList self);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_log_native.h

@@ -0,0 +1,25 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_log_slow(at::TensorList self);
+TORCH_API void _foreach_log_out(at::TensorList self, at::TensorList out);
+TORCH_API void foreach_tensor_log_slow_(at::TensorList self);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_log_cuda(at::TensorList self);
+TORCH_API void foreach_tensor_log_cuda_(at::TensorList self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_max_cuda_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_max(at::TensorList self);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_max_native.h

@@ -0,0 +1,23 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_max_slow(at::TensorList self);
+TORCH_API void _foreach_max_out(at::TensorList self, at::TensorList out);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_max_cuda(at::TensorList self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_neg_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,26 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_neg(at::TensorList self);
+TORCH_API void _foreach_neg_out(at::TensorList out, at::TensorList self);
+TORCH_API void _foreach_neg_outf(at::TensorList self, at::TensorList out);
+TORCH_API void _foreach_neg_(at::TensorList self);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_norm_native.h

@@ -0,0 +1,23 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_norm_slow(at::TensorList self, const at::Scalar & ord=2, ::std::optional<at::ScalarType> dtype=::std::nullopt);
+TORCH_API void _foreach_norm_Scalar_out(at::TensorList self, const at::Scalar & ord, ::std::optional<at::ScalarType> dtype, at::TensorList out);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_norm_cuda(at::TensorList self, const at::Scalar & ord=2, ::std::optional<at::ScalarType> dtype=::std::nullopt);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_pow_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,35 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(at::TensorList self, at::TensorList exponent);
+TORCH_API void _foreach_pow_out(at::TensorList out, at::TensorList self, at::TensorList exponent);
+TORCH_API void _foreach_pow_outf(at::TensorList self, at::TensorList exponent, at::TensorList out);
+TORCH_API void _foreach_pow_(at::TensorList self, at::TensorList exponent);
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(at::TensorList self, const at::Scalar & exponent);
+TORCH_API void _foreach_pow_out(at::TensorList out, at::TensorList self, const at::Scalar & exponent);
+TORCH_API void _foreach_pow_outf(at::TensorList self, const at::Scalar & exponent, at::TensorList out);
+TORCH_API void _foreach_pow_(at::TensorList self, const at::Scalar & exponent);
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(at::TensorList self, at::ArrayRef<at::Scalar> exponent);
+TORCH_API void _foreach_pow_out(at::TensorList out, at::TensorList self, at::ArrayRef<at::Scalar> exponent);
+TORCH_API void _foreach_pow_outf(at::TensorList self, at::ArrayRef<at::Scalar> exponent, at::TensorList out);
+TORCH_API void _foreach_pow_(at::TensorList self, at::ArrayRef<at::Scalar> exponent);
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(const at::Scalar & self, at::TensorList exponent);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_pow_cuda_dispatch.h

@@ -0,0 +1,29 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(at::TensorList self, at::TensorList exponent);
+TORCH_API void _foreach_pow_(at::TensorList self, at::TensorList exponent);
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(at::TensorList self, const at::Scalar & exponent);
+TORCH_API void _foreach_pow_(at::TensorList self, const at::Scalar & exponent);
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(at::TensorList self, at::ArrayRef<at::Scalar> exponent);
+TORCH_API void _foreach_pow_(at::TensorList self, at::ArrayRef<at::Scalar> exponent);
+TORCH_API ::std::vector<at::Tensor> _foreach_pow(const at::Scalar & self, at::TensorList exponent);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_sin.h

@@ -0,0 +1,44 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Function.h
+
+#include <ATen/Context.h>
+#include <ATen/DeviceGuard.h>
+#include <ATen/TensorUtils.h>
+#include <ATen/TracerMode.h>
+#include <ATen/core/Generator.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+
+
+
+#include <ATen/ops/_foreach_sin_ops.h>
+
+namespace at {
+
+
+// aten::_foreach_sin(Tensor[] self) -> Tensor[]
+inline ::std::vector<at::Tensor> _foreach_sin(at::TensorList self) {
+    return at::_ops::_foreach_sin::call(self);
+}
+
+// aten::_foreach_sin_(Tensor(a!)[] self) -> ()
+inline void _foreach_sin_(at::TensorList self) {
+    return at::_ops::_foreach_sin_::call(self);
+}
+
+// aten::_foreach_sin.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_sin_out(at::TensorList out, at::TensorList self) {
+    return at::_ops::_foreach_sin_out::call(self, out);
+}
+// aten::_foreach_sin.out(Tensor[] self, *, Tensor(a!)[] out) -> ()
+inline void _foreach_sin_outf(at::TensorList self, at::TensorList out) {
+    return at::_ops::_foreach_sin_out::call(self, out);
+}
+
+}

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_sinh_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_sinh {
+  using schema = ::std::vector<at::Tensor> (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_sinh";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_sinh(Tensor[] self) -> Tensor[]";
+  static ::std::vector<at::Tensor> call(at::TensorList self);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_sinh_ {
+  using schema = void (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_sinh_";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_sinh_(Tensor(a!)[] self) -> ()";
+  static void call(at::TensorList self);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_sinh_out {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_sinh";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_foreach_sinh.out(Tensor[] self, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_sqrt_native.h

@@ -0,0 +1,25 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_sqrt_slow(at::TensorList self);
+TORCH_API void _foreach_sqrt_out(at::TensorList self, at::TensorList out);
+TORCH_API void foreach_tensor_sqrt_slow_(at::TensorList self);
+TORCH_API ::std::vector<at::Tensor> foreach_tensor_sqrt_cuda(at::TensorList self);
+TORCH_API void foreach_tensor_sqrt_cuda_(at::TensorList self);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_foreach_zero_ops.h

@@ -0,0 +1,50 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _foreach_zero_ {
+  using schema = void (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_zero_";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_zero_(Tensor(a!)[] self) -> ()";
+  static void call(at::TensorList self);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+struct TORCH_API _foreach_zero_out {
+  using schema = void (at::TensorList, at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_zero";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_foreach_zero.out(Tensor[] self, *, Tensor(a!)[] out) -> ()";
+  static void call(at::TensorList self, at::TensorList out);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self, at::TensorList out);
+};
+
+struct TORCH_API _foreach_zero {
+  using schema = ::std::vector<at::Tensor> (at::TensorList);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_foreach_zero";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_foreach_zero(Tensor[] self) -> Tensor[] self_out";
+  static ::std::vector<at::Tensor> call(at::TensorList self);
+  static ::std::vector<at::Tensor> redispatch(c10::DispatchKeySet dispatchKeySet, at::TensorList self);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_functional_assert_scalar_ops.h

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _functional_assert_scalar {
+  using schema = at::Tensor (const at::Scalar &, c10::string_view, const at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_functional_assert_scalar";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_functional_assert_scalar(Scalar self, str assert_msg, Tensor dep_token) -> Tensor";
+  static at::Tensor call(const at::Scalar & self, c10::string_view assert_msg, const at::Tensor & dep_token);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Scalar & self, c10::string_view assert_msg, const at::Tensor & dep_token);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_fused_dropout_cuda_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor> _fused_dropout(const at::Tensor & self, double p, ::std::optional<at::Generator> generator=::std::nullopt);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_fused_moving_avg_obs_fq_helper_cpu_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cpu {
+
+TORCH_API ::std::tuple<at::Tensor,at::Tensor> _fused_moving_avg_obs_fq_helper(const at::Tensor & self, const at::Tensor & observer_on, const at::Tensor & fake_quant_on, at::Tensor & running_min, at::Tensor & running_max, at::Tensor & scale, at::Tensor & zero_point, double averaging_const, int64_t quant_min, int64_t quant_max, int64_t ch_axis, bool per_row_fake_quant=false, bool symmetric_quant=false);
+
+} // namespace cpu
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_fused_sdp_choice_native.h

@@ -0,0 +1,23 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/core/Tensor.h>
+#include <tuple>
+#include <vector>
+
+
+namespace at {
+namespace native {
+TORCH_API int64_t _fused_sdp_choice_cpp(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & attn_mask={}, double dropout_p=0.0, bool is_causal=false, ::std::optional<double> scale=::std::nullopt, bool enable_gqa=false);
+TORCH_API int64_t _fused_sdp_choice_cuda(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & attn_mask={}, double dropout_p=0.0, bool is_causal=false, ::std::optional<double> scale=::std::nullopt, bool enable_gqa=false);
+TORCH_API int64_t _fused_sdp_choice_meta(const at::Tensor & query, const at::Tensor & key, const at::Tensor & value, const ::std::optional<at::Tensor> & attn_mask={}, double dropout_p=0.0, bool is_causal=false, ::std::optional<double> scale=::std::nullopt, bool enable_gqa=false);
+} // namespace native
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_fw_primal_compositeexplicitautograd_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace compositeexplicitautograd {
+
+TORCH_API at::Tensor _fw_primal(const at::Tensor & self, int64_t level);
+
+} // namespace compositeexplicitautograd
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_histogramdd_from_bin_cts_ops.h

@@ -0,0 +1,39 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _histogramdd_from_bin_cts {
+  using schema = at::Tensor (const at::Tensor &, at::IntArrayRef, ::std::optional<at::ArrayRef<double>>, const ::std::optional<at::Tensor> &, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_histogramdd_from_bin_cts";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_histogramdd_from_bin_cts(Tensor self, int[] bins, *, float[]? range=None, Tensor? weight=None, bool density=False) -> Tensor";
+  static at::Tensor call(const at::Tensor & self, at::IntArrayRef bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+  static at::Tensor redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::IntArrayRef bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density);
+};
+
+struct TORCH_API _histogramdd_from_bin_cts_out {
+  using schema = at::Tensor & (const at::Tensor &, at::IntArrayRef, ::std::optional<at::ArrayRef<double>>, const ::std::optional<at::Tensor> &, bool, at::Tensor &);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_histogramdd_from_bin_cts";
+  static constexpr const char* overload_name = "out";
+  static constexpr const char* schema_str = "_histogramdd_from_bin_cts.out(Tensor self, int[] bins, *, float[]? range=None, Tensor? weight=None, bool density=False, Tensor(a!) out) -> Tensor(a!)";
+  static at::Tensor & call(const at::Tensor & self, at::IntArrayRef bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density, at::Tensor & out);
+  static at::Tensor & redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & self, at::IntArrayRef bins, ::std::optional<at::ArrayRef<double>> range, const ::std::optional<at::Tensor> & weight, bool density, at::Tensor & out);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_histogramdd_from_bin_tensors_cpu_dispatch.h

@@ -0,0 +1,23 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cpu {
+
+TORCH_API at::Tensor _histogramdd_from_bin_tensors(const at::Tensor & self, at::TensorList bins, const ::std::optional<at::Tensor> & weight={}, bool density=false);
+
+} // namespace cpu
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_int_mm_cuda_dispatch.h

@@ -0,0 +1,25 @@
+#pragma once
+// @generated by torchgen/gen.py from DispatchKeyFunction.h
+
+// NB: The implementing C++ file is RegisterDispatchKey.cpp
+
+// The only #includes we need are for custom classes that have defaults in the C++ API
+#include <c10/core/MemoryFormat.h>
+#include <c10/core/Scalar.h>
+#include <ATen/core/Reduction.h>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+
+namespace cuda {
+
+TORCH_API at::Tensor _int_mm(const at::Tensor & self, const at::Tensor & mat2);
+TORCH_API at::Tensor & _int_mm_out(at::Tensor & out, const at::Tensor & self, const at::Tensor & mat2);
+TORCH_API at::Tensor & _int_mm_outf(const at::Tensor & self, const at::Tensor & mat2, at::Tensor & out);
+
+} // namespace cuda
+} // namespace at

lib/python3.12/site-packages/torch/include/ATen/ops/_linalg_check_errors_ops.h

@@ -0,0 +1,28 @@
+#pragma once
+
+// @generated by torchgen/gen.py from Operator.h
+
+#include <tuple>
+#include <vector>
+
+// Forward declarations of any types needed in the operator signatures.
+// We can't directly include these classes because it will cause circular include dependencies.
+// This file is included by TensorBody.h, which defines the Tensor class.
+#include <ATen/core/ATen_fwd.h>
+
+namespace at {
+namespace _ops {
+
+
+struct TORCH_API _linalg_check_errors {
+  using schema = void (const at::Tensor &, c10::string_view, bool);
+  using ptr_schema = schema*;
+  // See Note [static constexpr char* members for windows NVCC]
+  static constexpr const char* name = "aten::_linalg_check_errors";
+  static constexpr const char* overload_name = "";
+  static constexpr const char* schema_str = "_linalg_check_errors(Tensor info, str api_name, *, bool is_matrix) -> ()";
+  static void call(const at::Tensor & info, c10::string_view api_name, bool is_matrix);
+  static void redispatch(c10::DispatchKeySet dispatchKeySet, const at::Tensor & info, c10::string_view api_name, bool is_matrix);
+};
+
+}} // namespace at::_ops

lib/python3.12/site-packages/torch/include/ATen/ops/_log_softmax_backward_data_meta.h

@@ -0,0 +1,27 @@
+#pragma once
+
+// @generated by torchgen/gen.py from NativeMetaFunction.h
+
+#include <c10/core/Scalar.h>
+#include <c10/core/Storage.h>
+#include <c10/core/TensorOptions.h>
+#include <c10/util/Deprecated.h>
+#include <optional>
+#include <c10/core/QScheme.h>
+#include <ATen/core/Reduction.h>
+#include <ATen/TensorIterator.h>
+#include <ATen/TensorMeta.h>
+#include <tuple>
+#include <vector>
+
+namespace at {
+namespace meta {
+
+struct TORCH_API structured__log_softmax_backward_data : public at::impl::MetaBase {
+    
+    
+    void meta(const at::Tensor & grad_output, const at::Tensor & output, int64_t dim, at::ScalarType input_dtype);
+};
+
+} // namespace native
+} // namespace at